Lecture Notes

On
Second Order Partial Differential
Equations
Outline
o Linear PDE with constant coefficients
o Introduction
o Method of Solving Linear PDE
 Operator method
 Separation method
 Fourier Transform method
 1.1 INTRODUCTION
 A second order linear PDE in two independent variables 𝑥 and 𝑦
can be written as
𝐴𝑢𝑥𝑥 + 𝐵𝑢𝑥𝑦 + 𝐶𝑢𝑦𝑦 + 𝐷𝑢𝑥 + 𝐸𝑢𝑦 + 𝐹𝑢 = 𝐺(𝑥, 𝑦) ---------- (1)
where 𝐴, 𝐵 and 𝐶 cannot be all zero and 𝑢 = 𝑢(𝑥, 𝑦) and 𝐴, 𝐵, 𝐶,
𝐷, 𝐸, 𝐹, 𝐺 are functions of 𝑥 and 𝑦 only.
 The second order PDE (1) is also classified as elliptic, parabolic or
hyperbolic, at any point (𝑥, 𝑦), depending on the value of the
discriminant
𝐵2 − 4𝐴𝐶 ------------------------------------ (2)
 The second order PDE (1) is:
1. elliptic if 𝐵2 − 4𝐴𝐶 < 0
2. hyperbolic if 𝐵2 − 4𝐴𝐶 > 0
3. parabolic if 𝐵2 − 4𝐴𝐶 = 0
Example: Classify the following Second Order PDE

2 2 𝑦2 𝑦2
1. 𝑦 𝑢𝑥𝑥 − 2𝑥𝑦𝑢𝑥𝑦 + 𝑥 𝑢𝑦𝑦 = 𝑢 + 𝑢𝑦
𝑥 𝑥 𝑦

Solution: 𝐴 = 𝑦 2 , 𝐵 = −2𝑥𝑦, 𝐶 = 𝑥 2
⟹ 𝐵2 − 4𝐴𝐶 = 4𝑥 2 𝑦 2 − 4𝑥 2 𝑦 2 = 0
Therefore, the given equation is Parabolic.
2. 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0.
𝐴 = 1. 𝐵 = 0, 𝐶 = 0 ⟹ 𝐵2 − 4𝐴𝐶 = −4 < 0
 Therefore, the given equation is elliptic
 The PDE 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 is called Laplace equation
𝜕2 𝑢 2𝜕 𝑢
2
3. = 𝑐
𝜕𝑡 2 𝜕𝑥 2
𝐴 = 𝑐 2 , 𝐵 = 0, 𝐶 = −1. ⟹ 𝐵2 − 4𝐴𝐶 = 4𝑐 2 > 0.
Therefore, the given equation is hyperbolic.
𝜕2 𝑢 2𝜕 𝑢
2
 The PDE = 𝑐 is called wave equation.
𝜕𝑡 2 𝜕𝑥 2
𝜕𝑢 𝜕 𝑢2
4. = 𝑐2 2. ⟹ 𝐴 = 𝑐 2 , 𝐵 = 0, 𝐶 = 0. ⟹ 𝐵2 − 4𝐴𝐶 = 0
𝜕𝑡 𝜕𝑥
Therefore, the given equation is parabolic.
𝜕𝑢 2
𝜕 𝑢
 The PDE = 𝑐2 2 is called one-dimensional heat equation.
𝜕𝑡 𝜕𝑥

𝜕𝑢 2 𝜕2 𝑢 𝜕2 𝑢
 The PDE = 𝑐 + is called two-dimensional heat equation.
𝜕𝑡 𝜕𝑥 2 𝜕𝑦 2
 Operator Method
 The linear partial differential equations with constant
coefficients can be written in the form
𝜕 𝜕
𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 𝑓(𝑥, 𝑦), where = 𝐷𝑥 , = 𝐷𝑦 ------(1)
𝜕𝑥 𝜕𝑦
 Example:𝑢𝑥𝑥 − 𝑢𝑦𝑦 + 2𝑢𝑥 + 𝑢 = 𝑥 2 + 𝑥𝑦
2 − 𝐷 2 + 2𝐷 + 1 𝑢 = 𝑥 2 + 𝑥𝑦
Operator form: 𝐷𝑥𝑥 𝑦𝑦 𝑥
 Complementary functions (solution) (C.F) of (1)
 The general solution of the homogenous linear PDE:
𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 0
 Particular integral (P.I) of (1)
 Any particular solution of (1) which contains no arbitrary
constants or functions.
• General integral (Solution) of (1)
𝑢 = C. F + P. I
Theorem: If 𝑢1 , 𝑢2 , … , 𝑢𝑛 are solutions of the
homogenous PDE: 𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 0, then
𝑛
𝑟<0 𝑐𝑟 𝑢𝑟 , where 𝑐𝑟 ′ 𝑠 are arbitrary constants is also a
solution.
Definition: The operator (1) is said to be
 reducible if it can be factorized into linear factors of
the type: 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 , where 𝛼, 𝛿 and 𝛽 are
constants.
 irreducible if it is not reducible
 Theorem: If 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 is a factor of
𝐹(𝐷𝑥 , 𝐷𝑦 ) and 𝜙𝑟 (𝑧) is an arbitrary function of
single variable 𝑧, then
−𝛿𝑟 𝑥
𝑢𝑟 = 𝜙𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 𝑒 𝛼𝑟 , 𝛼𝑟 ≠ 0 -----------(3)
is a solution of 𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 0
 Definition: A PDE 𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 𝑓(𝑥, 𝑦) is said
to be
 reducible if 𝐹(𝐷𝑥 , 𝐷𝑦 ) can be written as a product
of linear factors in 𝐷𝑥 and 𝐷𝑦 .
𝑛
That is 𝐹 𝐷𝑥 , 𝐷𝑦 = 𝑟<0 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟
 irreducible if it is not reducible.
 Solution of Reducible Equations
1. Case of no repeated factors:
Consider a reducible PDE
𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 𝑓(𝑥, 𝑦)
𝑛
Then 𝐹 𝐷𝑥 , 𝐷𝑦 = 𝑟<0 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟
 If 𝑢𝑟 satisfies
𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 𝑢𝑟 = 0, 𝑟 = 0,1,2, … , 𝑛 – (4)
then it gives us the complementary function.
 From (4), we have
𝑑𝑥 𝑑𝑦 𝑑𝑢𝑟
= = -------- (5)
𝛼𝑟 𝛽𝑟 ;𝛿𝑟 𝑢𝑟
Then, from (5) we obtain
𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 = 𝑐𝑟
−𝛿𝑟 𝑥 ,
𝑢𝑟 = 𝐴𝑟 𝑒 , 𝛼𝑟 ≠ 0
𝛼𝑟

Where 𝑐𝑟 and 𝐴𝑟 are arbitrary constants

−𝛿𝑟 𝑥
⟹ 𝑢𝑟 = 𝜙𝑟 𝑐𝑟 𝑒 𝛼𝑟 , 𝐴𝑟 = 𝜙𝑟 𝑐𝑟 for arbitrary
function 𝜙𝑟 .
 −𝛿𝑟 𝑥
Therefore, 𝐶. 𝐹 = 𝑛𝑟<1 𝜙𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 𝑒 𝛼𝑟 , 𝛼𝑟 ≠ 0
Particular case: 𝛼𝑟 = 0
𝑛
;𝛿𝑟 𝑦
𝐶. 𝐹 = 𝜙𝑟 𝛽𝑟 𝑥 𝑒 𝛽𝑟
𝑟<1
2. Case of Repeated Factors:
Suppose the PDE
𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 𝑓 𝑥, 𝑦
2
has repeated factors, say 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 .
2
Then 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 𝑢 = 0
⟹ 𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 𝑢 = 𝑢1 , where
𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 𝑢1 = 0
;𝛿𝑟 𝑥
𝑢1 = 𝜙𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 𝑒 𝛼𝑟
Therefore,
−𝛿𝑟 𝑥
𝛼𝑟 𝐷𝑥 + 𝛽𝑟 𝐷𝑦 + 𝛿𝑟 𝑢 = 𝜙𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 𝑒 𝛼𝑟 -----(6)
 From(6), we have
𝑑𝑥 𝑑𝑦 𝑑𝑢
= = −𝛿𝑟 𝑥 --------(7)
𝛼𝑟 𝛽𝑟
𝜙𝑟 𝛽𝑟 𝑥;𝛼𝑟 𝑦 𝑒 𝛼𝑟 ;𝛿𝑟 𝑢
 It follows from (7) that
𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 = 𝑐𝑟 -------- (8)
−𝛿𝑟 𝑥
𝑑𝑥 𝛿𝑟 1
+ = 𝜙𝑟 𝑐𝑟 𝑒 𝛼𝑟 -------(9)
𝛼𝑟 𝛼𝑟 𝑢 𝛼𝑟
 Integrating factor of equation (9) is
 𝛿𝑟 𝑥
𝐼. 𝐹 = 𝑒 𝛼𝑟

 Then the solution becomes

−𝛿𝑟 𝑥
𝑥
𝑢 𝑥, 𝑦 = 𝑒 𝛼𝑟 𝜙𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 + 𝜓𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 ,
𝛼𝑟
𝛼𝑟 ≠ 0
 Remark:
 This procedure can be generalized up to any order of
repetition of factors.
 Adding this solution to the sum of the other solutions
corresponding to the linear factors without repetition, we
get the required complementary function.
 Rules for finding the C.F
𝑛 𝑛;1 ′ ′𝑛
𝐷 + 𝑘1 𝐷 𝐷 + ⋯ + 𝑘𝑛 𝐷 𝑢 = 𝑓(𝑥, 𝑦)

By putting 𝐷 = 𝐷𝑥 = 𝑚, 𝐷′ = 𝐷𝑦 = 1
∴ Auxiliary equation is:
𝑚𝑛 + 𝑘1 𝑚𝑛;1 + ⋯ + 𝑘1 = 0. Let its roots be
𝑚1 , 𝑚2 , … all different, then
 C.F= 𝑓1 𝑦 + 𝑚1 𝑥 + 𝑓2 𝑦 + 𝑚2 𝑥 + ⋯
 If the roots 𝑚1 = 𝑚2 = 𝑚 (say),that is, two roots are
equal, then C.F= 𝑥𝜙 𝑦 + 𝑚𝑥 + 𝜓(𝑦 + 𝑚𝑥)
 When a root 𝑚 is repeated 𝑟 times, the
corresponding part of the C.F would be
 𝑓1 𝑦 + 𝑚𝑥 + 𝑥𝑓2 𝑦 + 𝑚𝑥 + 𝑥 2 𝑓3 𝑦 + 𝑚𝑥 +
⋯ + 𝑥 𝑟;1 𝑓𝑟 𝑦 + 𝑚𝑥

𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
 Example: Solve 2 2 + 5 + 2 2 =0
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑥

Solution: The equation in symbolic form can be

written as
2𝐷 2 + 5𝐷𝐷 ′ + 2𝐷 ′2 𝑢 = 0 ------ (1)
 Put 𝐷 = 𝑚, 𝐷 ′ = 1 in (1)
Hence its auxiliary equation: 2𝑚2 + 5𝑚 + 2 = 0
1
⟹ 𝑚 = − , −2.
2
 The general solution of (1) is given by
1
𝑢 𝑥, 𝑦 = 𝑓 𝑦 − 𝑥 + 𝑔(𝑦 − 2𝑥)
2
 Example: Solve

a) 𝐷 3 − 6𝐷 2 𝐷 ′ + 11𝐷𝐷 ′2 − 6𝐷 ′3 𝑢 = 0

b) (𝐷 4 −𝐷′4 )𝑢 = 0
Rules for Finding Particular Solution
 The particular solution of
𝐹 𝐷𝑥 , 𝐷𝑦 𝑢 = 𝑓(𝑥, 𝑦) is given by
1
𝑃. 𝐼(𝑃. 𝑆) = 𝑓(𝑥, 𝑦)
𝐹 𝐷𝑥 , 𝐷𝑦
Cases:
𝑎𝑥:𝑏𝑦 1
1. when 𝑓 𝑥, 𝑦 = 𝑒 ,then 𝑃. 𝐼 = 𝑒 𝑎𝑥:𝑏𝑦
𝐹 𝐷𝑥 ,𝐷𝑦
1 𝑎𝑥:𝑏𝑦 1
= 𝑒 , ≠0
𝐹 𝑎,𝑏 𝐹 𝑎,𝑏
1
 If = 0, then it is the case of failure.
𝐹 𝑎,𝑏
2. When 𝑓 𝑥, 𝑦 = sin 𝑎𝑥 + 𝑏𝑦 or cos(𝑎𝑥 + 𝑏𝑦)
1
𝑃. 𝐼 = 2 2
sin 𝑎𝑥 + 𝑏𝑦
𝐹 𝐷𝑥 , 𝐷𝑥 𝐷𝑦 , 𝐷𝑦
1
= sin 𝑎𝑥 + 𝑏𝑦 ,
𝐹 ;𝑎2 ,;𝑎𝑏,;𝑏2
1
2 2
≠0
𝐹 −𝑎 , −𝑎𝑏, −𝑏
Otherwise it is called a case of failure.
1
Note: To evaluate 𝑃. 𝐼 = 𝑓(𝑥, 𝑦), we resolve
𝐹 𝐷,𝐷′
1
into partial fractions treating 𝐹 𝐷, 𝐷′ as a
𝐹 𝐷,𝐷′
function of 𝐷 alone and operate each partial fraction
On 𝑓 𝑥, 𝑦 , remembering that
1
′
𝑓 𝑥, 𝑦 = 𝑓 𝑥, 𝑐 − 𝑚𝑥 𝑑𝑥
𝐷 − 𝑚𝐷

Where 𝑐 is replaced by 𝑦 + 𝑚𝑥 after integration.

Example: 1. Solve the equation

𝑢𝑥𝑥𝑥 − 2𝑢𝑥𝑥𝑦 − 𝑢𝑥𝑦𝑦 + 2𝑢𝑦𝑦𝑦 = 𝑒 𝑥:𝑦

Solution: The given equation can be rewritten as

𝐷𝑥3 − 2𝐷𝑥2 𝐷𝑦 − 𝐷𝑥 𝐷𝑦𝑦

2
+ 2𝐷𝑦3 𝑢 = 𝑒 𝑥:𝑦 --- (10)
 𝐹 𝐷𝑥 , 𝐷𝑦 = 𝐷𝑥3 − 2𝐷𝑥2 𝐷𝑦 − 𝐷𝑥 𝐷𝑦𝑦
2
+ 2𝐷𝑦3
= (𝐷𝑥 − 2𝐷𝑦 )(𝐷𝑥2 − 𝐷𝑦2 )
= (𝐷𝑥 − 𝐷𝑦 )(𝐷𝑥 + 𝐷𝑦 )(𝐷𝑥 − 2𝐷𝑦 )
Now 𝐷𝑥 − 2𝐷𝑦 𝑢 = 0 ⟹ 𝑢𝑥 − 2𝑢𝑦 = 0
𝑑𝑥 𝑑𝑦
⟹ = ⟹ 𝑦 + 2𝑥 = 𝑐1
1 −2
Thus any arbitrary function of 𝑦 + 2𝑥 say 𝑓1 (𝑦 + 2𝑥)
is a solution of equation (10)
𝑑𝑥 𝑑𝑦
Again 𝐷𝑥 + 𝐷𝑦 𝑢 = 0 ⟹ = ⟹ 𝑦 − 𝑥 = 𝑐2
1 1
Therefore, any arbitrary function of 𝑦 − 𝑥 say
𝑓2 𝑦 − 𝑥 is a solution of equation (10)
Similarly, 𝑓3 (𝑦 + 𝑥)is solution of equation (10)
corresponding to the factor 𝐷𝑥 − 𝐷𝑦 𝑢 = 0.
Thus, 𝐶. 𝐹 = 𝑓1 𝑦 + 2𝑥 + 𝑓2 𝑦 − 𝑥 + 𝑓3 (𝑦 + 𝑥)
Particular solution(P.I):
1
𝑃. 𝐼 = 𝑒 𝑥:𝑦
𝐹 𝐷𝑥 , 𝐷𝑦
1
= 𝑒 𝑥:𝑦
(𝐷𝑥 − 𝐷𝑦 )(𝐷𝑥 + 𝐷𝑦 )(𝐷𝑥 − 2𝐷𝑦 )
1 1
= 𝑒 𝑥:𝑦
𝐷𝑥 − 𝐷𝑦 (𝐷𝑥 + 𝐷𝑦 )(𝐷𝑥 − 2𝐷𝑦 )
1 1
= 𝑒 𝑥:𝑦
𝐷𝑥 − 𝐷𝑦 (1 + 1)(1 − 2)
 −1
= 𝑒 𝑥:𝑦
2(𝐷𝑥 −𝐷𝑦 )
1
Let 𝑧 = 𝑒 𝑥:𝑦 ⟹ 𝐷𝑥 − 𝐷𝑦 𝑧 = 𝑒 𝑥:𝑦
(𝐷𝑥 ;𝐷𝑦 )
𝑑𝑥 𝑑𝑦 𝑑𝑧
Therefore, = =
1 ;1 𝑒 𝑥+𝑦
First two members give us
𝑑𝑥 + 𝑑𝑦 = 0 ⟹ 𝑥 + 𝑦 = 𝑐 and
𝑑𝑧 𝑑𝑥 𝑐 = 𝑥𝑒 𝑥:𝑦
= ⟹ 𝑧 = 𝑥𝑒
𝑒𝑐 1
𝑧 1
Therefore, P.I= − = − 𝑥𝑒 𝑥:𝑦
2 2
Thus the complete solution is given by
𝑢 𝑥, 𝑦 = 𝑓1 𝑦 + 2𝑥 + 𝑓2 𝑦 − 𝑥 + 𝑓3 𝑦 + 𝑥
1 𝑥:𝑦
− 𝑥𝑒
2
• Example: Solve
𝜕2 𝑢 𝜕2 𝑢
• − = cos 𝑥 cos 2𝑦
𝜕𝑥 2 𝜕𝑥𝜕𝑦
• Solution: Writing the given equation in symbolic
form
• That is, 𝐷 2 − 𝐷𝐷′ 𝑢 = cos 𝑥 cos 2𝑦
• Its auxiliary is (by putting 𝐷 = 𝑚, 𝐷 ′ = 1)
𝑚2 − 𝑚 = 0
⟹ 𝑚 = 0,1
∴ C.F = 𝑓1 𝑦 + 0. 𝑥 + 𝑓2 𝑦 + 𝑥 = 𝑓1 𝑦 +
𝑓2 𝑦 + 𝑥
1
Now P.I= cos 𝑥 cos 2𝑦
𝐷2 ;𝐷𝐷′
 1 1
• = cos 𝑥 + 2𝑦 + cos(𝑥 − 2𝑦
2 𝐷2 ;𝐷𝐷′
1 1 1
= 2
cos 𝑥 + 2𝑦 + 2 cos(𝑥 − 2𝑦
2 𝐷 − 𝐷𝐷′ 𝐷 − 𝐷𝐷′
Put 𝐷 2 = −1, 𝐷𝐷 ′ = −2 in first term
𝐷 2 = −1, 𝐷𝐷 ′ = 2 in second term
1 1 1
= cos 𝑥 + 2𝑦 + cos(𝑥 − 2𝑦
2 −1 + 2 −1 − 2
1 1
= cos 𝑥 + 2𝑦 − cos(𝑥 − 2𝑦
2 6
Hence complete solution is:
1
𝑢 𝑥, 𝑦 = 𝑓1 𝑦 + 𝑓2 𝑦 + 𝑥 + cos 𝑥 + 2𝑦 −
2
1
cos(𝑥 − 2𝑦
6
 SEPARATION OF VARIABLES METHOD
 The method of separation of variables is applicable
to a large number of classical linear homogeneous
equations. The choice of the coordinate system in
general depends on the shape of the body.
 Consider the Laplace equation:
𝛻 2 𝑢 = 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 ---------(1)
 Assume that 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦) ------(2)
𝑋 ′′ 𝑌 ′′
 Equations (1) and (2) provide us = − =𝑘
𝑋 𝑌
(separation parameter) we have three cases.
Cases 1: Let 𝑘 > 0. Then 𝑘 = 𝑝2 , where 𝑝 is a real
𝑑2𝑋 𝑑2𝑇
number, we get − 𝑝2 𝑋 = 0 and + 𝑝2 𝑇 = 0
𝑑𝑥 2 𝑑𝑦 2
, which implies that
𝑋 = 𝑐1 𝑒 𝑝𝑥 + 𝑐2 𝑒 ;𝑝𝑥 and 𝑇 = 𝑐3 cos 𝑝𝑦 + 𝑐4 sin 𝑝𝑦.
Thus the solution is
𝑢 𝑥, 𝑦 = 𝑐1 𝑒 𝑝𝑥 + 𝑐2 𝑒 ;𝑝𝑥 𝑐3 cos 𝑝𝑦 + 𝑐4 sin 𝑝𝑦
𝑑2 𝑋 𝑑2𝑇
Case 2: Let 𝑘 = 0.Then = 0 and = 0 which
𝑑𝑥 2 𝑑𝑦 2
Provide us X = 𝑐5 𝑥 + 𝑐6 and 𝑌 = 𝑐7 𝑦 + 𝑐8
Therefore, the solution is
𝑢 𝑥, 𝑦 = 𝑐5 𝑥 + 𝑐6 𝑐7 𝑦 + 𝑐8 .
 Case 3: Let 𝑘 < 0. Then 𝑘 = −𝑝2 , proceeding as in the
case 1, we get
𝑢 𝑥, 𝑦 = 𝑐9 cos 𝑝𝑥 + 𝑐10 sin 𝑝𝑥 𝑐11 𝑒 𝑝𝑦 + 𝑐12 𝑒 ;𝑝𝑦
 In all these cases 𝑐𝑖 𝑖 = 1, … , 12 are integration
constants which are calculated by using the boundary
conditions.

 For example, consider the boundary conditions

𝑢 𝑥, 0 = 0, 𝑢 𝑥, 𝑎 = 0, 𝑢(𝑥, 𝑦) ⟶ 0, where

𝑥 ≥ 0 and 0 ≤ 𝑦 ≤ 𝑎
 The appropriate solution for 𝑢(𝑥, 𝑦) by the method
of separation of variables obtained above in this case
is 𝑢 𝑥, 𝑦 = 𝑐1 𝑒 𝑝𝑥 + 𝑐2 𝑒 ;𝑝𝑥 𝑐3 cos 𝑝𝑦 + 𝑐4 sin 𝑝𝑦
 Since 𝑢(𝑥, 𝑦) ⟶ 0, as 𝑥 ⟶ ∞, we have 𝑐1 = 0 for all 𝑦.

 Therefore, 𝑢 𝑥, 𝑦 = 𝑐2 𝑒 ;𝑝𝑥 𝑐3 cos 𝑝𝑦 + 𝑐4 sin 𝑝𝑦

 As 𝑢 𝑥, 0 = 0, we get 𝑐2 𝑒 ;𝑝𝑥 𝑐3 = 0 ⟹ 𝑐3 = 0
since 𝑐2 ≠ 0 ≠ 𝑒 ;𝑝𝑥 for all 𝑥.
 Therefore, 𝑢 𝑥, 𝑦 = 𝐴𝑒 ;𝑝𝑥 sin 𝑝𝑦, where 𝐴 = 𝑐2 𝑐4
 Now, 𝑢 𝑥, 𝑎 = 0 ⟹ 𝐴𝑒 ;𝑝𝑥 sin 𝑝𝑎 = 0
⟹ sin 𝑝𝑎 = 0, since 𝐴 ≠ 0. Then 𝑝𝑎 = 𝑛𝜋, 𝑛 ∈ ℤ.
 𝑛𝜋
 ⟹𝑝= ,𝑛 = 0, ±1, ±2, ±3 …
𝑎
−𝑛𝜋
∞ 𝑛𝜋
 Therefore, 𝑢 𝑥, 𝑦 = 𝑛<0 𝐴𝑛 𝑒 𝑎 sin 𝑦, 𝐴𝑛
𝑎
being new constant. This is the required solution in
this case.
 FOURIER TRANSFORM
 Criteria for choosing F.S.T.:
1. At least one of the independent variables should
have the range from 0 to ∞ and apply F.S.T. w.r.t.
that variable only.
2. The value of the unknown function 𝑢 must
be known at the lower limit of the variable which
has the range from 0 to ∞.
𝜕𝑢
3. The behavior of 𝑢(𝑥, 𝑡) and at 𝑥 = ∞ should
𝜕𝑥
𝜕𝑢
be known (if not applied, 𝑢 and ⟶ 0 as
𝜕𝑥
𝑥⟶∞
 Criteria for choosing F.C.T.:
1. At least one of the independent variables should
have the range from 0 to ∞ and apply F.C.T. w.r.t.
that variable only.
𝜕𝑢
2. The value of the unknown function must
𝜕𝑥
be known at the lower limit of the variable which
has the range from 0 to ∞.
𝜕𝑢
3. The behavior of 𝑢(𝑥, 𝑡) and at 𝑥 = ∞ should
𝜕𝑥
𝜕𝑢
be known (if not applied, 𝑢 and ⟶0 as
𝜕𝑥
𝑥⟶∞
 Criteria for choosing general Fourier Transform (F.T)
1. At least one of the independent variables should
have the range from −∞ to ∞ and apply F.T. w.r.t.
that variable only.
𝜕𝑢
2. Both 𝑢 and must vanish as 𝑥 ⟶ ±∞
𝜕𝑥
 𝜕𝑢 𝜕2 𝑢
Example 1: = 𝑘 2 , 0 < 𝑥 < ∞, with 𝑢 𝑥, 0 = 0
𝜕𝑡 𝜕𝑥
𝜕𝑢
when 𝑥 > 0, = −𝜇 (constant) when 𝑥 = 0, and 𝑢,
𝜕𝑥
𝜕𝑢
⟶ 0, as 𝑥 ⟶ ∞, 𝑡 > 0.
𝜕𝑥
Solution:Taking the Fourier cosine Transform of both
sides of the given equation (w.r.t. 𝑥), we get

2 ∞ 𝜕𝑢 2 ∞ 𝜕2 𝑢
0
cos 𝛼𝑥𝑑𝑥 =𝑘 0 2 cos 𝛼𝑥𝑑𝑥 --- (1)
𝜋 𝜕𝑡 𝜋 𝜕𝑥

∞
𝑑𝑢𝑐 2 𝜕2𝑢
⟹ =𝑘 2
cos 𝑛𝑥 𝑑𝑥
𝑑𝑡 𝜋 0 𝜕𝑥
 ∞ ∞
𝑑𝑢𝑐 2 𝜕𝑢 𝜕𝑢
⟹ =𝑘 cos 𝛼𝑥 +𝛼 sin 𝛼𝑥 𝑑𝑥
𝑑𝑡 𝜋 𝜕𝑡 0 0 𝜕𝑥
2 ∞ 𝜕𝑢 𝜕𝑢
=𝑘 𝜇+𝛼 0
sin 𝛼𝑥 𝑑𝑥 b/c ⟶0
𝜋 𝜕𝑥 𝜕𝑥
𝜕𝑢
as 𝑥 ⟶ ∞, = −𝜇 when 𝑥 = 0
𝜕𝑥
2 ∞ 2 ∞
=𝑘 𝜇 + 𝛼(𝑢 sin 𝛼𝑥) 0 −𝛼 0
𝑢 cos 𝛼𝑥 𝑑𝑥
𝜋

𝑑𝑢𝑐 2
⟹ = 𝑘𝜇 − 𝑘𝛼 2 𝑢𝑐 , since 𝑢 ⟶ 0 as 𝑥 ⟶ ∞
𝑑𝑡 𝜋

𝑑𝑢𝑐 2 2
Therefore, + 𝑘𝛼 𝑢𝑐 = 𝑘𝜇
𝑑𝑡 𝜋
 𝑘𝛼 2 𝑑𝑡 𝑘𝛼 2 𝑡
I.F= 𝑒 =𝑒
Therefore,

At 𝑡 = 0, 𝑢𝑐 = 0 (since 𝑢 = 0 at 𝑡 = 0)
2𝜇 2𝜇 ;𝑘𝛼 2 𝑡
∴𝐴=− 2
⟹ 𝑢𝑐 = 2
1−𝑒
𝜋𝛼 𝜋𝛼

2 ∞
∴ 𝑢 𝑥, 𝑡 = 𝑢𝑐 cos 𝛼𝑥 𝑑𝛼
𝜋 0
2𝜇 ∞ cos 𝛼𝑥 ;𝑘𝛼 2 𝑡
∴ 𝑢 𝑥, 𝑡 = 2
1−𝑒 𝑑α
𝜋 0 𝛼
 𝜕𝑢 𝜕2 𝑢
Example 2: = ,𝑥 > 0, 𝑡 > 0 subject to conditions
𝜕𝑡 𝜕𝑥 2
1,0 < 𝑥 < 1
𝑢 0, 𝑡 = 0, 𝑢 𝑥, 0 = and 𝑢(𝑥, 𝑡)
0, 𝑥 ≥ 1
is bounded
 Solution: Taking the Fourier sine Transform of both
sides of the given equation (w.r.t. 𝑥), we get

2 ∞ 𝜕𝑢 2 ∞ 𝜕2 𝑢
0
sin 𝛼𝑥𝑑𝑥 = 0 2 sin 𝛼𝑥𝑑𝑥 --- (1)
𝜋 𝜕𝑡 𝜋 𝜕𝑥

∞
𝑑𝑢𝑠 2 𝜕2𝑢
⟹ = 2
sin 𝑛𝑥 𝑑𝑥
𝑑𝑡 𝜋 0 𝜕𝑥
 𝑑𝑢𝑠 2 𝜕𝑢 ∞ ∞ 𝜕𝑢
• ⟹ = sin 𝛼𝑥 −𝛼 0
cos 𝛼𝑥 𝑑𝑥
𝑑𝑡 𝜋 𝜕𝑥 0 𝜕𝑥

2 ∞ 𝜕𝑢 𝜕𝑢
= −𝛼 0
cos 𝛼𝑥 𝑑𝑥 as ⟶0 as
𝜋 𝜕𝑥 𝜕𝑡
𝑥 ⟶ ∞.
up on integrating by parts, we have
∞
2 ∞
= −𝛼 (𝑢 cos 𝛼𝑥) − 𝛼2 𝑢 sin 𝛼𝑥 𝑑𝑥
𝜋 0 0

2 2 ∞
= −𝛼 0
𝑢 sin 𝛼𝑥 𝑑𝑥, as 𝑢 o, t = 0, and
𝜋
𝑢(𝑥, 𝑡) ⟶ 0 as 𝑥 ⟶ ∞.
 2 ∞
𝑢 sin 𝑛𝑥 𝑑𝑥 = 𝑢𝑠 .
𝜋 0
𝑑𝑢𝑠
⟹ = −𝛼 2 𝑢𝑠
𝑑𝑡
𝑑𝑢𝑠
Therefore, + 𝛼 2 𝑢𝑠 = 0 is first order ode.
𝑑𝑡
;𝛼 2 𝑡
𝑢𝑠 = 𝐴𝑒
2 ∞
At 𝑡 = 0, 𝑢𝑠 𝛼, 0 = 0
𝑢(𝑥, 0) sin 𝛼𝑥 𝑑𝑥
𝜋

2 1 2 ∞
= 0
1. sin 𝛼𝑥 𝑑𝑥+ 1
0. sin 𝛼𝑥 𝑑𝑥
𝜋 𝜋

2 cos 𝛼𝑥 1 2 1 − cos 𝛼
= − =
𝜋 𝛼 0 𝜋 𝛼
 2 1 − cos 𝛼
⟹ 𝐴=
𝜋 𝛼

2 1 − cos 𝛼 ;𝛼2 𝑡
⟹ 𝑢𝑠 = 𝑒
𝜋 𝛼
Now applying the inverse Fourier sine Transform, we
have

∞ ∞
2 2 1 − cos 𝛼 ;𝛼2 𝑡
𝑢 𝛼, 𝑡 = 𝑢𝑠 𝛼, 𝑡 sin 𝛼𝑥 𝑑𝛼 = 𝑒 sin 𝛼𝑥 𝑑𝛼
𝜋 0 𝜋 0 𝛼

 This is the required solution.